Cisco VP of AI Barak Turovsky on LLM Use Cases and IT Automation Benefits
For technology leaders, the strategic question is not whether to use LLMs, but where they belong in the automation stack. The value proposition is strongest when AI reduces friction in repeatable IT work: triage, knowledge retrieval, incident routing, service desk assistance and policy-aware guidance. That makes this less a science project and more a management decision about which workflows can tolerate probabilistic output and which still require deterministic controls.
That shifts the governance conversation. If an AI assistant can recommend actions, who approves the action, who owns the model output, and what audit trail is retained? CIOs and IT directors should treat LLMs as part of the operating model, not a standalone tool: define decision rights, human override points, escalation thresholds and acceptable-use boundaries before scaling beyond pilots.
The portfolio trade-off is equally important. Automation initiatives often compete for the same scarce capacity as infrastructure modernisation, security remediation and platform consolidation. Leaders should ask which use cases have the shortest path to measurable savings or service improvement, and how success will be tracked. Useful next questions include: what baseline metrics exist today, what counts as productivity gain, and where will any time saved actually be reinvested?
Finally, adoption will determine whether the technology creates trust or drag. Teams need clarity on when to rely on AI, when to verify it and how to report failures without penalty. The practical leadership task is to pair experimentation with workforce enablement, so automation strengthens service quality rather than creating hidden operational risk.
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.